This document discusses functional programming languages Scala and Clojure that run on the Java Virtual Machine (JVM). It provides an overview of what functional programming is, why it is becoming more important with multi-core processors, and why developing new languages on the JVM is advantageous due to its existing investments and performance. Specific details are given about Clojure, including how it has tight integration with Java and the JVM, uses immutable collections and lazy sequences, and provides primitives for concurrency. Scala is described as a functional language that is fast, expressive, and statically typed with features like traits, pattern matching, and XML literals. The document concludes that as developers we should choose the right tools for our

2. Functional Programming
on the JVM with
Scala and Clojure
What they are and why you should care
Cristiano Breuel Thadeu Russo
Advisory Software Engineer Advisory Software Engineer
(cbreuel@br.ibm.com) (thadeurc@br.ibm.com)

3. Y = λf.(λx.f (x x)) (λx.f (x x))

4. Y = λf.(λx.f (x x)) (λx.f (x x))

5. What is functional programming
about?

6. Lambda calculus (a.k.a λ-calculus)
function definition
function application
recursion

8. Referential Integrity
Avoids state and mutable data
Higher-Order functions:
Functions can receive and return other
functions
Lazy Evaluation:
Evaluate an expression only when
necessary

16. FP has been around since the 1950’s
Why it is becoming more important now?

17. FP has been around since the 1950’s
Why it is becoming more important now?

18. FP has been around since the 1950’s
Why it is becoming more important now?
Why should I care?

19. Because of them...

20. Because of them...
... the multi-core processors

21. Because of them...
... the multi-core processors
(also, to become a smarter
developer!)

22. On the other side...

23. On the other side...
The JVM is cool!

24. JVM -> Java Language
Huge investment already on the JVM
(libraries, middleware, performance, GC,
remote communication, etc..)
JVM executes byte-code

25. The question is:

26. The question is:
Why not create new (functional)
languages on top of the JVM?

28. (not Closure)

29. Clojure has...

30. Clojure has...
... tight integration with the JVM and Java

31. Clojure has...
... tight integration with the JVM and Java
... compilation for performance

32. Clojure has...
... tight integration with the JVM and Java
... compilation for performance
... dynamic evaluation

33. Clojure has...
... tight integration with the JVM and Java
... compilation for performance
... dynamic evaluation
... memory-efficient immutable collections

34. Clojure has...
... tight integration with the JVM and Java
... compilation for performance
... dynamic evaluation
... memory-efficient immutable collections
... lazy sequences

35. Clojure has...
... tight integration with the JVM and Java
... compilation for performance
... dynamic evaluation
... memory-efficient immutable collections
... lazy sequences
... primitives for concurrency (STM, agents)

36. Clojure is a LISP

37. Clojure is a LISP
(Lots of (Irritating
(Stupid (Parentheses))))

38. Clojure is a LISP
(Lots of (Irritating
(Stupid (Parentheses))))
a.k.a. List Processing

39. LISP

40. LISP
Homoiconicity - code is data

41. LISP
Homoiconicity - code is data
Macros

42. LISP
Homoiconicity - code is data
Macros
Multiple dispatch (multi-methods)

43. Syntax
Operands
Operator (arguments)
(str "Hello" "World")
Expression
delimiters

44. Syntax
["red" "green" "blue"] Array
{:color "blue" :length 2} Map
:color Keyword
#(even? %) Anonymous function

45. Higher-Order Functions

46. Efficient Immutable
Collections

47. Lazy Sequences
import java.util.HashMap;
import java.util.Map;
public class ColorMapper {
private static String[] COLORS = {"white", "blue"};
public static Map<String, String> mapLineColors(String[] values) {
Map<String, String> retVal = new HashMap<String, String>();
int i = 0;
for (String value: values) {
retVal.put(value, COLORS[i % COLORS.length]);
i++;
}
return retVal;
}
public static void main(String[] args) {
String[] values = {"Clojure", "Java", "Scala", "Visual Basic"};
System.out.println(mapLineColors(values));
}
}

48. Java Integration
(def now (new java.util.Date))
(def now (java.util.Date.))
(. now getTime)
(doto (JFrame.)
(. add (JLabel. "Hello, Clojure!"))
(. pack)
(. show))

49. sample

50. Scala
Fast
Functional
Expressive
Statically typed (with type inference)
Concurrent

51. every value is an object
vals / vars
traits
companion objects
currying
closures / high order functions
pattern matching / case classes
XML literals
Lazy evaluation, and much more...

52. sample

53. To sum up...

55. We cannot choose our problems....
.... but we can choose our tools.
It is our responsibility to figure out
the “right” tool.

56. Even if you can't use
these languages in your
day job yet, learning
them will make you a
better programmer!

57. References
http://clojure.org/
http://www.scala-lang.org/